Information Technology Reference
In-Depth Information
and/or supercomputers composed of significant numbers of computing units. Swin-
DeW-C is currently running at Swinburne University of Technology as a virtualized
environment which is physically composed of 10 servers and 10 high-end PCs at the
Swinburne Astrophysics Supercomputer Node. To simulate the cloud computing en-
vironment, we set up VMware [56] software on the physical servers and create virtual
clusters as datacenters. For more details, please refer to [27].
Fig. 2.
SwinDeW-C System Architecture [27]
As depicted in Figure 2, the architecture of SwinDeW-C is designed according to
classical WfMC reference model [57] . The four basic components of temporal verifi-
cation framework are implemented as part of the workflow QoS management tools in
SwinDeW-C. Specifically, the QoS management tools include QoS requirement
specification, QoS-aware service selection, QoS consistency monitoring and QoS
violation handling, and they are designed as generic and open components to meet
different QoS requirements such as time, cost, reliability, security and so on. For our
temporal verification framework, temporal constraint setting is part of QoS require-
ment specification and QoS-aware service selection which are mainly interacted with
the workflow modelling tool to create workflow specifications. The QoS requirement
specification component will generate the QoS constraints including the temporal
constraints, which are part of the workflow specification and the basic reference data
for QoS-aware service selection. The QoS-aware service selection component will
return the available (best and backup) software services satisfying the QoS con-
straints, through the broker service (part of the application provision services). After
the workflow specifications are submitted to the workflow enactment services, the
workflow instances can be executed by invoking software services which are ma-
naged by the tool agents. The tool agents will in charge for the data communication
between the workflow system and the software services which are either available in
the system local repository or delivered by outside service providers in the cloud mar-
ket. During runtime execution stage, the workflow execution state will be constantly
monitored by the QoS consistency monitoring component which includes temporal
